When Delhi was under colonial rule it suffered from an excess of venomous cobras. To curb the population the government paid a bounty for dead cobras. This triggered entrepreneurs to start breeding cobras to collect the bounty. When the government figured out what was happening, they discontinued the bounty which meant all the cobras being bred were worthless and were thus set free, increasing the cobra population significantly.

The Cobra effect is when the solution for a problem unintentionally makes the problem worse. And it happens more often than you might think. There are similar stories about incentives to kill rats in Hanoi, incentives to reduce greenhouse gasses, incentives to reduce narcotic production, and more.

I am also struck by the idea that incentives that sound terrible might actually produce good outcomes. I heard a story from a friend in South Africa that their town had legalized the hunting of endangered rhinoceroses. This sounds like a shockingly bad idea. But they instituted a very large fee that a hunter would have to pay that would be split with the farmer whose land the rhino was on. Farmers who used to tip off poachers to get rhinos off their land would now aggressively defend the rhino against poachers. I don’t know how widespread the program was but my friend suggested that at least in the first few years the population increased meaningfully.

There are more examples like this. The unsavory practice of keeping animals captive in zoos and aquariums has increased enthusiasm for protecting animals in the wild. Safe injection sites reduce negative externalities of illegal drug use. Sex education reduces incidences of teen pregnancy.

We get caught in these traps all the time as consumers. My favorite example is unlimited data cell phone plans. Everyone thinks they want an unlimited data plan. But at the incentive level the goal for the carrier there is to give you the minimum level of service required just so you won’t switch carriers. Instead if you have a plan where you pay for data their incentive is to make your experience using data as fast and as great as possible so you are inclined to use more.

I think of this all the time in the context of running a company. People propose things that seem like commonsense solutions that end up creating a ton of perverse incentives downstream. For example:

  • “everyone should share feedback directly instead of through managers” leads to people either withholding valuable feedback to maintain relationships or damaging relationships if they can’t share such feedback elegantly. This is a job the manager is trained to do.
  • “Let’s create a comprehensive description of what it takes to do each job at each level” leads to people only doing things that were written down rather than the common sense things that someone really at that level would naturally understand. The more comprehensive you attempt to make it the more trapped you are by it. The reason these things are vague is because all jobs of any meaning have tons of discretion and people are often only able to demonstrate they are at a certain level by how well they operate within that uncertainty. Almost by definition if you have to tell someone exactly what the job is then they aren’t there yet, outside of the most junior roles.
  • “Let’s cross calibrate against a wide set of other teams” when the teams don’t deeply understand one another’s work leads to a lowest common denominator understanding of value creation. Things like organizational size or how much press something got become important features when in practice making big impact with a small team or doing important work that nobody sees are likely much more important to emphasize.

As Sam Altman wrote in a recent blog post: “Incentives are superpowers; set them carefully.”